Robert Samuelson Oversells the Case for Economic Optimism

Last summer news reports were filled with ill-informed predictions of a double-dip recession. Now there seem to be many accounts that misrepresent recent economic data to make a case for substantially stronger growth.

Robert Samuelson makes some of the standard errors in outlining a case for optimism. (In fairness, the column also presents a case for pessimism.) For example, he touts the jump in housing starts reported for November, saying, "Housing construction was up 9.3 percent in November over October and 24.3 percent over November 2010."

The increase in starts reported in November was almost entirely attributable to a jump in starts reported for multi-family units. Multi-family starts are highly erratic and frequently have large month-to-month rises and falls. Starts of single-family homes were actually 1.5 percent below their November, 2010 level.

The piece also refers to a jump in pending home sales reported for November. This is a measure of contracts signed. The National Association of Realtors reports that many more contracts are now falling through than in the past, so this rise in contracts does not likely mean a corresponding rise in sales. (In this vein, purchase mortgage applications are running even with or below their year ago level.)

The column also notes that recent construction levels have been well below the number needed to keep even with household growth. This is correct, but we are still far from making up for the overbuilding of the bubble years as indicated by the fact that the vacancy rate remains at near record levels.

(There have been some questions raised about the accuracy of the Census Department's data, claiming that it overstates the number of housing units in the country. Those raising the issue fail to note that measures of housing starts do not include housing units that were created by conversion of commercial or industrial property, such as an old warehouse being turned into condos. The rehabilitation of dilapidated units would also not be included in housing start numbers. There were many cases of both ways of adding to the housing stock during the bubble years. Also, it is important to note that the Census data is giving the percentage of units that are vacant. The critics of this measure must show how the Census methodology would lead it to overstate the share of units that are vacant.)

Finally, the piece notes that household debt levels have fallen since the beginning of the recession, implying that there could be a consumption boom as families are now better positioned to make major purchases. While debt is down, so is wealth. Households have lost close to $8 trillion in housing wealth and another $4 trillion in stock wealth. This would be expected to lead to a sharp drop in consumption through the wealth effect.

At present the saving rate is close to 4.0 percent. This is considerably above the near zero rate at the peak of the bubble, but well below the 8.0 percent average of the pre-bubble years. It seems more likely that, given this massive loss of wealth, the savings rate would be more likely to rise than fall, especially with tens of millions of baby boomers approaching retirement with the prospect of having almost nothing other than their Social Security to support them.

I see several comments that refer to the 2010 Census data and imply that it shows fewer vacancies than the quarterly survey numbers that I have been using. I am not sure what 2010 Census data these comments are referring to, but the ones I see show a higher vacancy rate than the quarterly data.

Table 1 of the 2010 Census publication on housing characteristics reports a total of 131,705,000 housing units. This is somewhat higher than the 130,517,000 units reported in the quarterly survey for the second quarter of 2010, the period in which most data collection took place. This implies that the survey had been understating the number of housing units, not overstating as some previous comments had claimed.

The 2010 Census data showed that 116,716,000 of these units were occupied for an occupancy rate of 88.6 percent. This is somewhat lower than the 88.9 percent occupancy rate shown in the quarterly survey data (Table 3, combining full and part-year occupancy). In short, if the 2010 Census is showing us that the quarterly data is understating occupancy and overstating vacancies, I'm not finding it in this publication.

...
written by Chris, January 02, 2012 9:33 AM

Samuelson, according to Wikipedia, majored in government at Harvard. I would suspect he has no training as an economist. I wonder why the Post keeps using him to write about the economy since his abilities in the field appear insufficient for the job.

No Reason for Optimism Until Housing Prices Start Rising
written by Paul, January 02, 2012 11:59 AM

In every previous recession of the past 40 years, housing has led the recovery with rising sales and prices. Not this time. Housing prices have fallen for 5 straight years and continue to decline. We are just bouncing along the bottom of the Great Recession.

Samuelson
written by burt, January 02, 2012 12:04 PM

Samuelson is a journalist, not an academic. Saying than no one without a degree in economics can write about it is silly. We might as well say that only priests can write about morality.

...
written by S. D. Jeffries, January 02, 2012 2:21 PM

Samuelson may be a journalist and not an academic, but he has been writing columns on economics for years. Therefore he should have learned, at the very least, how to interpret a census report or take more than one perspective on a trade association's report to reach a conclusion. Samuelson has always been a journalist whose writing is more concerned with pushing an ideology than looking at all the available information, recognizing interactions among data sets, then letting a conclusion emerge organically.

...
written by Blissex, January 02, 2012 5:14 PM

«Households have lost close to $8 trillion in housing wealth and another $4 trillion in stock wealth.»

That's the usual colossal mistake. Absolutely no wealth is gained by anybody when house or share prices go up and none is lost when they go down.

The amount of wealth in the country is entirely unchanged: same number of houses, same number of plants and offices.

The only way to produce wealth is via production of goods and services, not via asset price bubbles (capital gains).

Economists *should* know that wealth changes only with GDP growth, with value added growth, and capital gains are not value added and don't change GDP.

What asset price inflation produces is purely upwards income redistribution via purchasing power generation, for example in a doubling of prices owners of $1 million houses get an extra $1 million of additional income via capital gains, and the owner of a $100 thousand house gets $100 thousand and non-owners get nothing extra. Since production of goods and services has not changed at all, the richer the owner, the more purchasing power he has been gifted with a capital gain, and the more he can buy outbidding the less rich owners or the poor.

Thus a fall in asset prices does not destroy any wealth, even if it impacts demand by decreasing purchasing power by canceling the effect of previous capital gains with capital losses.

-1

Someone tell Warren Buffet
written by AndrewS, January 02, 2012 6:38 PM

Poor guy thought he was wealthy all these years from accumulating all those capital gains.

HVS Data
written by JSeydl, January 02, 2012 7:09 PM

Dean,

Bill at CR responded to your comments about the HVS data: http://www.calculatedriskblog....s-and.html

You make a good point about conversions not being counted in the housing starts data. So if I understand this correctly, your claim is that the broader ACS and decennial data do not include conversion units, whereas the HVS data does, which is why the HVS data are showing a much larger number of vacant housing units? I wonder if we can get the U.S. Dept. Commerce to validate all of this. The difference between surveys is, after all, quite large. And that difference is very important for all of us trying to understand how long this housing slump will drag on!

CPS/HVS
written by Thomas Lawler, January 02, 2012 7:59 PM

The questions about the accuracy of the CPS/HVS data do not relate to housing units estimates, which of course the CPS/HVS does not estimate (it gets those estimates from another area of Census). The questions relate to the number and percent of housing units that are occupied vs. those that are vacant, as well as the type of vacant units. The CPS/HVS data are based on a small sample and is a voluntary survey. The decennial Census data, which show a much higher number (and %) of occupied units, is based on a much larger (duh) sample. There is compelling evidence that the CPS/HVS data (1) overstates significantly overall and rental vacancy rates; and (2) overstates significantly homeownership rates, especially among "younger" adults. There is also strong evidence that the "trends" shown in the CPS/HVS data on homeownership rates and vacancy rates are "off."

...
written by Calgacus, January 02, 2012 8:03 PM

Chris: Samuelson, according to Wikipedia, majored in government at Harvard. I would suspect he has no training as an economist.
Chris, these are positive credentials. Academic "economics" education of the last few decades, particularly at a center of nescience like Harvard, leaves the student more ignorant & stupid than when they entered - but not cognizant of the true value of their degree in "mathematical" homeopathic astrology. (Sorry, O homeopaths & astrologers). With the amount of training R. Samuelson has NOT gotten, one would expect him to do better though. His admitting an error deserves more praise, though, for its rarity.

Blissex, you're assuming ("outbidding") the economy is not demand constrained. It's not pure upward distribution. Dean is right, though perhaps telegraphic.

...
written by JSeydl, January 02, 2012 8:14 PM

Thanks, Thomas. Sounds like we should steer clear of the HVS data. This is unfortunate, since the decennial Census data come out only once every 10 yrs!

...
written by Michael Chase, January 03, 2012 1:19 AM

The documentation for the $8 trillion loss in housing wealth is evident, but where can I find documentation for the $4 trillion loss in stock wealth? Thank you.

...
written by JGBellHimself, January 03, 2012 3:39 AM

There are some minor, if you think so, things missing in Dean's comment.

While his point in Para 3 about single family starts is correct, he missed a far more significant fact, repeatedly pointed out by Bill@CR, namely that 2010 was the worst year on record for single family starts and sales = 323K. Worse, 2011 is now even less than that = about 305K.

If those numbers show U.S. getting better, we sure in Hades (AZ?) wouldn't want to see worse ones, either. So, let's all of U.S. simply not look at them.

In the 4th para, there are three problems:
FIRST, Steven Hansen @ Econintersect point out that even NAR's numbers simply do not support that "cancellations" argument - it simply is not happening.
See: http://econintersect.com/wordpress/?p=17389

SECOND, what everyone seems to disregard is the fact that NAR gets its "pending sales data" from the same local MLSs that they get there "existing home sales data". They just admitted that their EHS numbers were very badly overstated, and they revised them down. They did not, however, admit that PHS data suffers from the same problem(s).

THIRD, trying to compare mortgage apps to Pending homes sales is almost as irrelevant as it is impossible. While it might be true that there is a close relationship between new home sales and mortgage apps, bcuz few buyers pay cash - new home apps are not separated from existing sales apps, nor foreclosure purchase apps, nor short sales apps. Bcuz "all cash investors" are now a HUGE part of the foreclosure market - to include them in "existing home sales" is totally misleading.

Recently there has been both a huge decline in "foreclosure sales" to "all cash investor/buyers" AND a large decline in mortgage apps - so, where precisely are all these new, additional pending home buyers getting the money? Explain, please - NAR, or anyone?

Fourth, in para 5 the contention that "recent construction levels have been well below the number needed to keep even with household growth" is made. But, the author also points out the new construction is not single family homes, but "apartments".

There is a disconnect here. There is evidence, such as in a Seattle Times article, that we now are about to overbuild "apartments". What you do not factor in is that when condo sales stopped, they began to rent them out - to foreclosure evictions? Now, if they cannot rent all these new apartments, they can, can they not, simply sell them as condos.

And, that does not even consider the price to rent ratios that make buying a home less expensive than renting costs U.S. now.

FIFTH, in para 7 Dean points to the loss of "wealth" to suggest that to assume those people will rush back in to buy homes is misguided. What he either forgot or missed is that the loss of wealth, in de-leveraging was in two major areas: homes and credit.

With the massive lost equity in foreclosures and short sales and strategic defaults, many, if not most, of those people now can NOT buy any home - new or used. While some of U.S. have been paying down our mortgages, you did NOT mention how many of them (now 50%) are underwater (now by 50%) - and can neither sell nor buy.

With the possible exception of student loans and autos, there has been a huge decline in "other debt" - especially credit cards. However, regularly it is pointed out that was due to credit card issuers writing off and cutting off U.S. equal to the decline. While some did pay down, most were cut off. How many of them went through BKs? Will any of them qualify for a mortgage RE loan? How many years will they not be able to do that?

And, deleveraging does not even address the 20+% of U.S. who have lost our jobs, and cannot find a new one. Even if they do, will THEY qualify for a new mortgage loan, with that job history?

And, those who graduate with huge student loan obligations, will they qualify for a new mortgage loan - assuming they can, someday, even find a job?

...
written by JGBellHimself, January 03, 2012 3:44 AM

In paras 5 & 6 Dean attempts to address the problems with the Census "housing vacancy" numbers. Nice try, but no cookies.

For an in-depth analysis, and a disagreement with Dean Baker, you should see Bill@CR's and Lawler's comments on this article:
http://www.calculatedriskblog.com/2012/01/comments-on-housing-vacancies-and.html

Which states: "this report (HVS) is commonly used by analysts to estimate the excess vacant supply for housing, but - because the vacancy rates do not match the Census data (or the much larger ACS data) - it doesn't appear to be useful for that purpose.

However, as we have pointed out to Bill, the Census data is also very suspect, for a number of reasons.

For example, when the Census "counted" U.S. they did not ask if we were in default on our mortgages, and about to be foreclosed or evicted. Some will say that isn't important. In THIS foreclosure environment it isn't? Why?

Millions have been foreclosed on, and many more millions will shortly be foreclosed on, that the Census "assumes" were not and are not now vacant.

Similarly, with the huge increase in job losses, and the large increase in household "consolidations" - kids moving back with parents and parents moving in with kids - did either the 2010 Census, or the CVS survey, OR the current projections of the changes now taking place, assume that this is, or that is it not, happening?

As Bill point out the CVS is "benchmarked" to the 2010 Census. But, does the way they do that match what changed in the few years before the Census was taken, or to reflect the changes that are still taking place? Or, to they make assumptions based upon the differences between 2000 and 2010? They are very different.

The Census just issued new population "estimates" that say AZ is 9th in new growth. How was that computed - using decennial adjustments, or does it adjust for the last few years. What AZ did between 2000 and 2010 is VERY different from what happened between 08 and 11. Which was used, a decennial growth adjustment, or the recent contraction adjustment?

Are both the Census (along with the projections based on it) and the CVS numbers as inaccurate as NAR's numbers? Why should we assume they are not.

...
written by JGBellHimself, January 03, 2012 3:48 AM

Oh, and oddly, there is a much better measure - current and an accurate count.

As we pointed out to Bill@CR, worrying about how many angels among U.S. are dancing on the head of a pin - assuming that they really ARE angels...

making their mortgage/rent payments, and not illegal aliens in a drop house.

why not use..., drum roll, please:

The US Postal Service count of "deliverable homes"...
that they provide each and every month to "bulk mailers".

Vacancies Data
written by JSeydl, January 03, 2012 9:40 AM

I think you need to look at Q4 2010 in the HVS for an apples to apples comparison. Here's what I'm finding.

HVS
Q4 2010 vacancy rate = 2.60%
Q4 2010 total housing stock = 131,975,000
Therefore, Q4 2010 vacancies = 3.43 million

ACS
2010 vacancy rate = 2.49%
2010 total housing stock = 131,791,065
Therefore, 2010 vacancies = 3.28 million

Decennial Census
2010 vacancy rate = 2.40%
2010 total housing stock = 131,704,730
Therefore, 2010 vacancies = 3.16 million

So it looks like the quarterly HVS is overstating vacancies, but not by much (only by about 150K relative to the ACS at the end of 2010). I thought the discrepencies were supposed to be in the millions???

-Joe

...
written by Blissex, January 03, 2012 12:13 PM

«Someone tell Warren Buffet
written by AndrewS, January 02, 2012 7:38 PM
Poor guy thought he was wealthy all these years from accumulating all those capital gains.»

He has accumulated earnings not mere capital gains, and I am surprised that there are people who don't understand the difference. Real estate owners have accumulated nothing, just seen price changes.

If you really believe that capital gains are wealth, the government could simply enact a law that house prices to double each year to make all house proprietors ever wealthier and America the wealthiest country forever, and all this for free and at no cost to anyone. Pareto optimality into infinity!

it is pure upward redistribution
written by Blissex, January 03, 2012 12:18 PM

«Blissex, you're assuming ("outbidding") the economy is not demand constrained. It's not pure upward distribution.»

A widespread increase in real estate prices does not by itself increase production of anything, so it is simply putting more purchasing power into the hands of the proprietors of bigger properties.

It might eventually lead to an increase in the construction of houses, but that's just the effect of price inflation (that is handing out free purchasing power to real estate proprietors).

Just throwing dollar bills from helicopters would be a fairer and more effective way to support demand than giving free capital gains to bigger proprietors.

Robert Samuelson Oversells the Case for Economic Optimism

About Beat the Press